From: bbense+comp.lang.ruby.Apr.05.02@... Date: 2002-04-06T01:34:22+09:00 Subject: Re: a note on the . and .. discussion -----BEGIN PGP SIGNED MESSAGE----- In article <3b3ad3b4.0204022320.7225f2c6@posting.google.com>, Patrick May wrote: >Yohanes Santoso wrote in message news:<871ydxaj7m.fsf@jenny-gnome.dyndns.org>... >> patrick-may@monmouth.com (Patrick May) writes: >> > >Does the mapping of "." to the current directory happen on the >filesystem level, i.e. the directories hash of names to inodes >contains a "." and a ".." entry? > - - Yes, . and .. are in the directory entry. Some shells however override the behavior of .. to avoid "confusing" the user. If you work in big complicated filesystems with lot's of symlinks there can be quite a difference between head( current_working_directory ) and ... - - Some shells will override .. and replace it with head( current_working_directory ). For example, suppose my home dir is listed as /usr/users/bbense in the password file, but is actually a symlink to a directory in a shared filesystem (like AFS ), i.ed. /afs/ir/users/b/b/bbense - - When I log in my $cwd will be /usr/users/bbense - - However .. in that directory will be /afs/ir/users/b/b/ But in some shells if I type cd .. The shell will override the meaning of .. and actually run this command cd /usr/users - - Booker C. Bense -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBPK3PTQD83u1ILnWNAQG8uAP/QnbkAGo+tUb7HqbHWoNa968yUK6vesWz 63hsewd/4vLyhF1YFeMJs0PsOzpQ2iqPDsGhIvPF/VQ0+1A7SlLRJU4hyXyFSK8K gGB7pBK/b+/IP79GHd5heNGFSYKrcecUfWXEmOKScRQtx126HgUDFRotz+L/xkAL U9aMfl1FQoA= =g7co -----END PGP SIGNATURE----- --